Telling stories of terrorism: a framework for applying narrative approaches to the study of militant's self-accounts

ABSTRACT Narrative has recently garnered in much attention in the study of terrorism but remains poorly understood. This paper offers some initial steps towards translating the promise of narrative approaches into a set of steps for systematically analysing and understanding terrorists’ own accounts of their engagement with extremism and militancy. This approach rests on the assumption that terrorist authored accounts are more than post-hoc rhetorical exercises that aim to persuade others, or even the authors themselves, of the righteousness of their political cause or otherwise mitigate their responsibility for their involvement in violence. In particular, I advance a framework for methodically applying narrative approaches to terrorist authored texts, in particular, autobiographies. In doing so, I will demonstrate how this approach can help better comprehend how individuals involved in militancy understand the world, draw upon existing narrative resources and give meaning to their actions.

. A good example of this lack of precision is demonstrated within a single line of the United Kingdom's House of Commons Home Affairs Committee report (2016) which demands 'an urgent need for strong narratives to counter the evil but effective messages which terrorist organisations are disseminating' (p. 5). Disappointingly, much academic work falls into the same trap, exhibiting a similarly cavalier attitude towards the term 'narrative' whilst reproducing many of the same unsubstantiated assumptions upon which policy-focused research rests (see for example -Leuprecht, Hataley, Moskalenko, & McCauley, 2009McDowell-Smith, Speckhard, & Yayla, 2017). In both the practitioner and academic spheres then, narrative is frequently taken to be synonymous with a message, belief, worldview or ideology. A narrative, however, is none of these things. As Glazzard (2017) argues, when it comes to narrative, this theoretical dearth is likely a result of the conspicuous absence of literary studies from terrorism research. In particular, he convincingly demonstrates how the failure to properly understand and study terrorist communications as narratives in a literary sense undermines efforts to comprehend their appeal. I argue, however, that Glazzard's assertion extends further than merely analysis of terrorist propaganda and is also crucial for interrogating terrorists' self-accounts.
Although terrorism studies has for the large part lagged behind the rest of the social sciences in embracing a theoretically grounded appreciation of narrative and narrative approaches, recently there has been something of a wave of interest in giving narrative a bigger platform within the field. Pemberton and Aarten's (2017), assertion that they 'hope to have ignited interest in cross-fertilization between narrative approaches and the topic of terrorism and political violence' (p. 12), may demonstrate a lack of awareness of some of the existing research within and on the periphery of terrorism studies that has employed narrative approaches, but is nevertheless a welcome sentiment. 1 Significantly, much of this new wave of narrative-grounded research has looked to move beyond merely analysing terrorists' efforts to communicate and has sought to highlight the place of narrative in understanding individuals' personal experiences of engaging in political violence. The recent special issue of Studies in Conflict and Terrorism edited by Graef, da Silva and Lemay-Hebert not only makes a case for the use of narrative approaches more generally in terrorism research but also for analysing the accounts of those who have engaged in political violence (2018). Significantly, in proposing narrative as a 'tool' for researching political violence they recognise and pose the question that has hampered much research in this area, namely the explanation of 'what concrete steps do they [researchers] take to analyze stories? ' (2018, p. 7). Problematically, however, this same criticism could be levelled at some of their own work. In particular, da Silva's previous work provides little explanation of her method of analysis, and at points, appears to repeat the stories of her subjects uncritically (2017). Relatedly, within the same special issue, Madina, Bilbao, and Bermudez (2018) make a critically important contribution on the educational capacity of fictional narratives that regard political violence. However, their assertion that the capacity of fictional worlds created in literary narrative to influence individuals' real-life beliefs has not yet been explored appears to entirely overlook important research in literary studies, most notably Green and Brock's work on transportation into narrative worlds, that demonstrates exactly this (2000,2002). Whilst not to denigrate their approach, this omission is nevertheless indicative of the disciplinary grounding of much of this new research in firmly psychological, rather than literary understandings of narrative. 2 Different theoretical approaches are to be expected somewhat given that narrative, like terrorism studies, sits within different disciplinary spaces. Nevertheless, Glazzard's criticism regarding the absence of literary studies appears, at least to some extent, to continue to ring true. The result is a lack of appreciation for some of the novel theoretical and methodological approaches that exist elsewhere (see for example Bearman & Stovel, 2000).
Research from other disciplines has, however, successfully brought different understandings of narrative together. In particular, the demonstrated applications of relevant theoretical literature, as well as methodological processes, in narrative criminology provide a good template for scholars of terrorism to aspire to (see for example Presser & Sandberg, 2015bsee also Colvin & Pisoiu, 2018). In following their lead this paper aims to demonstrate how different disciplinary insights on narrative can be brought to bear in studying the accounts of terrorists as the literary texts that they are, as well as their capacity as windows into the psychological processes of their authors. In doing so, the aim is to avoid the methodological shortcomings in existing studies and offer theoretical clarity to enhance further research into narrative accounts of terrorism. As such, I will outline a framework for approaching and analysing terrorist authored accounts (primarily autobiographical works) that demonstrates how relevant theoretical insights can be mapped directly onto these texts in a manner that is easily replicable for analysts and researchers. The first part of this paper is dedicated to laying out the theoretical basis for the approach taken; exploring questions of what narrative is, why it is important in the move to violence and the impact it can have. The second part will show how the methodological framework works in practice, drawing attention to questions of narrative imagination and transportation. Finally, I will argue that paratextual issues, along with three aspects of narrative discoursegenre, time and coherencecan be fruitfully employed to understand and analyse terrorism.

What is narrative?
Given that this article has been critical of other research for failing to define 'narrative', it is necessary to take a step back and add some conceptual clarity to the object of study. First, it is important to stress that narrative is ubiquitous to human existence, being both transhistorical and transcultural (Barthes, 1982, pp. 251-252). Despite this, narrative remains something of a contested concept. No single, commonly accepted definition exists even in literary theory. There are, however, some central features and common definitional themes of narrative that are widely agreed upon. It is commonly accepted that for something to constitute a narrative, an event, however abstract, must take place. In this sense, unlike a chronology (or a simple list of dates) narratives necessarily depict a change from one condition to another (Brockmeier, 1993;Bruner, 1990;Ricoeur, 1984). In this sense, narrative is a uniquely human phenomenon, occurring in all activities that serve to represent events in time (Porter Abbott, 2002, p. Xi). Significantly, narrative allows events or incidents themselves to create the order of time, with this understanding resting on the assumption that two events necessarily belong together and logically follow one another. In this way narrative is more than merely description; it not only illustrates but explains (Polletta, 2006;Squire et al., 2014). In doing so, narrative unavoidably generates meaning, making a moral point, no matter how subtle (Bruner, 1990;Polletta, Chen, Gharrity Gardner, & Motes, 2011;Polkinghorne, 1988;White, 1980). Some narrative researchers make a distinction between 'stories' as merely a recounted sequence of events, and 'narratives' as accounts of events that require some level of organisation, plotting and interpretation on behalf of the narrator (Chatman, 1975, p. 295). Such a distinction can be traced, in part, to the Russian structuralist linguists of the 1920s who, in establishing contemporary narratology, insisted on the distinction between story or 'fabula', and narrative or 'syuzhet' (see for example -Propp, 1968). By this understanding, a story is never seen directly but instead is picked up through the 'narrative discourse', or events as represented (Porter Abbott, 2002, p. 14). This implies that a necessary level of symbolic work is undertaken on the behalf of the author in presenting an event or sequence of events as they took place (Chatman, 1975, p. 296). Approaching the analysis of narratives in such a manner, it is possible to contrast the 'what' of 'stories' (or the content) with the 'how and why' of 'narratives' (their form, structure and the context in which they are told).
Although the structuralist separation of story and narrative is not without problems, it nevertheless provides a useful starting point for the discussion of narrative analysis presented within this paper. In the first instance, this division is valuable in interrogating the limitations of much of the existing research that has sought to apply narrative approaches to the study of terrorist authored accounts. It is clear that the majority of research, even some presented under the banner of narrative grounded approaches, has focused on the content of terrorist self-narratives, at the expense of analysing their form. For the most part, analysis of these texts has in part focused on identifying and coding the occurrence of biographic incidents and episodes (see for example Altier, Horgan, & Thoroughgood, 2012). Alternatively, as Cottee and Haywood highlight, autobiographical texts have been examined in light of what they reveal about how terrorists think ideologically (2011, p. 946). In both cases, this has led to much analysis that, whilst valuable in its own right is, as Glazzard (2017) states, 'one-dimensional, limited to an identification of recurring motifs and themes without attending to the texture or technique of terrorist narrative' (p. 9). Similarly, this approach fails to consider the place of storytelling and what stories do as well as what they reveal. As highlighted, however, this paper argues that the form and the purpose of stories, are as crucial, if not more so than their content, in understanding individuals' experiences of involvement in militancy.
The story/narrative divide can serve a valuable organisational function in the analytical breakdown of narrative texts. However, an important caveat must be made here; namely whether such a neat division between the two can, in fact, be made. Whenever an account of a sequence of events is presented it is inevitably narrativised because any attempt to do so is necessarily inflected by the storyteller's narrative culture (Squire et al., 2014). This is something especially apparent when individuals come to provide accounts of their own experiences. Faced with a limitless array of events, characters and experiences culture inevitably influences the prioritisation and selection of certain incidents over others (Somers, 1994, p. 617). Even subtle influences of culture, such as notions of what 'makes a good story', blur the line between 'story' and 'narrative' to the point where it becomes elusive and open to subjective interpretation. In this light, many researchers reject this division, especially given the tendency of structurally-based approaches to decontextualise stories from the circumstances of their telling (Culler, 2002). Furthermore, given that this paper advocates an approach that focuses on the place of storytelling as well as the stories themselves, the analytic utility of the story/narrative division becomes questionable. Nevertheless, in putting aside theoretical concerns, this division will be used operationally throughout this paper to help demonstrate how the breaking down of texts can be undertaken systematically and as a useful means to illustrate some of the theoretical considerations discussed. 3

Rethinking narrative and experience
The story/narrative divide is useful for exploring one of the central challenges that has hampered the study of terrorists' self-accounts. In conceiving narrative discourse as something separate to story, and responsible for, explaining events as they transpired, this structuralist division then makes a claim about the relationship between narrative and experience; namely that story necessarily pre-exists narrative. The underlying assumption is that human lives, as lived, exist 'out there' somewhere, independent of narrative description (Presser, 2010, p. 434). In this sense, narrative provides either an objectivethat is accurate description of events and the world as it really isor a subjective, personal interpretation of these incidents as they happened and the world they occurred within (Presser, 2010, p. 434). Concisely, narrative then serves either as a record or interpretation of experience. In viewing narrative in such manner, the utility of using self-accounts for the purposes of research quickly becomes questionable. As Hopkins argues, 'the perennial question of truthfulness and authenticity in memoir-writing has coloured the social scientific reception for this type of source material' (Hopkins, 2013, p. 8). Such fears have dominated research on political violence to the extent that autobiographical accounts written by current and former terrorists have received relatively little systematic analysis despite long standing recognition of their potential as a source of detailed information (Altier et al., 2012;Cordes, 1987;Rapoport, 1987).
Despite the apparently inherent logic that narrative follows experience, it is important to highlight that this is only one way of conceiving the interplay between the two. Here the work of French philosopher Paul Ricoeur is crucial (1984). Ricoeur conceptualises the relationship between experience and narrative as dynamic: 'aspects of experience itself are presented originally as they appear in the narration and that narrative form is not simply imposed on preexistent real experiences but helps to give them form' (Polkinghorne, 1988, pp. 67-68). In this sense, narratives are much more than post hoc justifications but rather are constitutive of, or shape, experience (Presser, 2010, p. 434). Given that we experience the world narratively, as the protagonist of our own ever-changing story, it becomes clear that 'a life is not "how it was" but how it is interpreted and reinterpreted, told and retold' (Bruner, 1987, p. 31see also Fleetwood, 2015;Presser, 2009, p. 184). This position also circumvents an inherent tension in viewing narrative as a record or interpretation of experience; the expectation that an individual's story will necessarily be the same in each retelling despite the dynamic nature of human life (Polletta, 2006, p. 3). It is logical that, as our experiences are constantly changing and expanding, so too must our narratives of how we understand these events and the world (Presser, 2010, p. 434). Again, thinking about self-narratives in this way supports the notion that stories do something that is they are always told for different purposes or audiencesand this inevitably influences how they are narrated (Presser & Sandberg, 2015a, p. 3).
This understanding of narrative and experience forms the backbone of the sub-discipline of narrative criminology, which has sought to analyse the role of stories in individuals' engagement in and understanding of acts of harm (Maruna, 2001;Presser, 2009;Presser & Sandberg, 2015b). Of particular relevance this work has included analysis of the accounts of those who have engaged in politically motivated violence (Berntzen & Sandberg, 2014;Sandberg, 2013Sandberg, , 2015. Somewhat surprisingly, this conceptualisation also finds explicit support within the autobiographies of a number of terrorists. Former M-19 guerrilla, Vásquez Perdomo (2005), for example, in the preface to her autobiography, describes how she discovered meaning in the process of recalling her life. She states, 'the written words and I influenced each other, we affected each other always. Thanks to this exercise, I found meanings and explanations that had been invisible to me' (pp. Xxxiii-xxxiiv).

Narrative, meaning and action
As Vásquez Perdomo's statement highlights, in conceptualising narrative and experience as inseparable, the importance of meaning comes to the fore. It is a widely accepted that humans innately seek meaning and purpose in life (King, Heintzelman, & Ward, 2016). Whilst establishing the link between meaning and action sits well beyond the scope of this paper, examining the theoretical underpinnings of this relationship is nevertheless necessary to properly understand the centrality of narrative in the move between the two.
Psychology has long focused on discovering and describing meaning-making at the level of the individual, or the symbolic activities that humans employ in constructing and making sense of both the world and themselves (Bruner, 1990, p. 2). However, given that meaning is neither biologically determined nor something created solely in the mind of the individual, its study has always also necessitated insight from interpretive disciplines in the humanities and social sciences (Bruner, 1990, p. 2;Mattingly, Lutkehaus, & Throop, 2008, p. 2). Meanings, instead, are collective and communal; the product of cooperative creation and negotiation between individuals (Bruner, 1990, pp. 12-13). Meaning-making, a central impulse of human life, then necessarily constitutes a narrative process. Stories draw on the events that matter in our lives, incessantly replacing meaningsvalues, judgements, motivations, commitments and emotionsin place of a straightforward chronological recollection of the events recounted (Barthes, 1977, p. 119;Presser, 2010, p. 431). Through narrative we are able to bring our own meanings to the public domain, further renegotiating and reconstituting them. Storytelling, then, is meaning-making; in other words, individuals do not merely express meaning through stories but rather fundamentally create meaning in the process of constituting their experiences in narrative form (Bruner, 1990;Gottschall, 2012). However, despite their public nature, meanings are not always easily identified. The dialogic nature of how meanings are created and recreated, contested and resisted, means that meaning is inherently unstable, forever in flux, even where it appears constant (Andrews, 2014, p. 3;Bohman, 1997, p. 176). The fluid nature of language itself, is reflected in the multiplicity that meanings can also take. Similarly, meanings are not always fully appreciated by those who hold them, with individuals only ever having a partial understanding as to the meanings of the stories they tell (Presser, 2010, p. 444;Ricoeur, 1976).
Meanings do not in themselves, however, constitute restraints or triggers for action. Instead, meanings constitute a 'possibility relationship' between individuals and the worlds they inhabit, signalling or indicating a range of options or possibilities for action (Brockmeier, 2009, p. 222see also Holzkamp, 1983, p. 236). Each individual is born into a world where culture and language provide pre-existing 'symbolic systems' that provide the tools necessary to undertake the unescapable task of having to interpret, weigh and ultimately choose between established cultural meanings (Brockmeier, 2009, p. 222;Bruner, 1990). As Somers (1994) argues, 'people are guided to act in certain ways, and not others, on the basis of the projections, expectations, and memories derived from a multiplicity but ultimately limited repertoire of available social, public, and cultural narratives' (p. 614). Similarly, Bruner states, 'culture and the quest for meaning within culture are the proper causes of human action ' (1990, p. 20). To return this discussion to political violence, narratives endow certain readings of events with local understandings of coherence and emotional meaning (Polletta, 2006, p. 53;Tölölyan, 1989, p. 101). As social movement scholars have argued, these narratives then often provide rationales for participation in collective action and can induce within individuals the feeling of a moral obligation to act (Polletta, 2006, p. 6). Crucially, such narratives can not only 'imbue the time of individual lives with transcendent collective values', but also 'tell individuals how they would ideally have to live and die in order to contribute properly to their collectivity and its future' (Tölölyan, 1989, p. 101).

Narrative imagination and transportation to narrative worlds
If cultural meanings are accepted as being inescapable, questions regarding the place of individuals in shaping their own meanings arise. Despite the importance of culture, individuals always have a say in negotiating meaning. In this light, narratives that help fashion meaning should be thought of as reflective of both structure and agency; simultaneously reflective of existing structures, since social positioning and experience make available particular kinds of discourses but at the same time a consequence of individual imagination (Fleetwood, 2015;Presser & Sandberg, 2015b;Sandberg & Pederson, 2009). Imagination is an under theorised but nevertheless central element of political violence. The pursuit of all political goals, even those advanced by violence, is inherently linked to imagination, specifically the ability to imagine alternative realities, both in terms of our own lives and the world we live in. This somewhat disrupts everyday assumptions regarding the division of the real and the imaginary. However, as Andrews (2014) argues, 'the real and the not-real are not then polar opposites but, rather, are positioned in relation to one another, linked by a thread of ongoing change and perpetual becoming' (p. 6). Narrative, again, is key here. Crucially, the progression from a present situation to a hoped for but not yet realised alternative reality is always rendered in narrative form. Simply, before any action can take place, a narrative emplotment, which includes characters, action and a desired endpoint, must first be arrived at in the imagination of the individual (Andrews, 2014, p. 5;Ringmar, 1996, p. 73). Such narrative imagination is therefore a necessary process that proceeds terrorists' actions from the execution of a single act of violence to their wider aims of conceiving and realising alternative political realities. 4 Imagination also raises questions about individuals' capacity and willingness to not only embrace narratives but, further, to immerse themselves within them and the alternative realities they constitute. Whilst this paper has primarily focused on terrorists as producers of literature, it is also important to recognise these individuals are also consumers of narrative texts (Glazzard, 2017 see also Berger, 2016;Holbrook, 2017). Holbrook found in his study of the media consumption of a number of convicted terrorists in the UK that stories, fables and biographical accounts conveyed as heroic narratives were especially prevalent (2017, p. 28). Berger has similarly documented the wide ranging link and influence of the post-apocalyptic and racist novel, the Turner Diaries, to various white nationalists who have engaged in violence (2016).
Again, literary studies offers valuable insight into the affective capacity of such texts. Notably, Green and Brock (2002) have focused on the power of narrative to stimulate within individuals 'transportation into narrative worlds', or 'a state in which a reader becomes absorbed in the narrative world, leaving the real world, at least momentarily behind' (p. 317see also Green, 2004;Green & Brock, 2000). In a rather simplified sense, most people can recall being 'lost' in a story at some point or another (Nell, 1988). Narrative's unique power to bring about such immersion is interesting for a number of reasons. First, the traveller (the reader, listener, viewer or receiver of the narrative) always returns from the journey of transportation somewhat changed by the experience (Gerrig, 1993, pp. 10-11). Commonly, transportation induces an emotional change in the real-world state of the traveller, creating an affective bridge that connects this narrative world to the everyday one (Green, Brock, & Kaufman, 2004, p. 317). Similarly, transportation has the power to activate and link previously held images in the mind of those transported and, furthermore, produce new vivid images that are intrinsically bound to the wider narrative advanced by the text (Green & Brock, 2002). Crucially, the result is that transportation into narrative worlds invariably influences individuals' real-world views upon their return (Green, 2004;Green & Brock, 2000;Green et al., 2004). Interestingly, transportation is not affected by whether the receiver knew the narrative to be fictive or true (Green & Brock, 2000. Furthermore, a number of studies indicate that fictive narratives can bring about changes in beliefs as much, if not more so, than factual ones (Green, 2004;Green & Brock, 2000;Slater, 1990;Strange & Leung, 1999). The traveller may know that these events did not happen but can still nevertheless experience real emotions and believe that similar events could take place in the real world in future (Green, 2004, p. 251). 5 Again, returning to Berger's study (2016), the Turner Diaries appears to have taken on just such capacity for a number of white nationalists who have possessed copies of the book and have gone on to commit acts of violence (p. 31). As he notes, that the book is written as an epistolary novelor that which present themselves in the form of a series of 'found documents' such as letters, diary entries or news clippingadds to this text's realism (2016, p. 14). This, along with its dystopian theme, also seeks to induce a sense of urgency that something must be done to avoid such a future. In this sense, the Turner Diaries displays one way in which the division between 'real' and 'fictional' worlds can begin to become disrupted.
Incidents of transportation are also explicitly mentioned in a number of terrorist autobiographies. For many, like Danish born Jihadi, Morten Storm,who in describing first reading a book about the life of the Prophet Mohammed states 'within minutes I was so absorbed in the story that the world outside evaporated ' (2014, p. 24) and that he continued to read the book for six hours without break (2014, p. 26)transportation serves to demonstrate the impact of certain texts in their ideological development. Furthermore, an indication of transportation to narrative worlds is often the 'illusion of intimacy' between those transported and the inhabitants of these fictional worlds (Horton & Wohl, 1956). Here even the distinction between a reader and a narrative's characters (usually protagonists) often becomes disrupted. In this sense, those transported absorb and become these characters, which in turn may then have a real-world impact on their beliefs and actions. This is influenced by the reader's experiences and perceptions of the characters in these worlds sharing similar backgrounds and traits to their own (Green, 2004). Vásquez Perdomo's autobiography (2005) details her experience as a youngster of playing the role of Adela in a dramatic performance of Garcia Lorca's play La cas de Bernada Alba; 'this was the key that opened the door to another world for me … the role of Adela in La cas de Bernada Alba changed my life ' (pp. 20-21). Immersing herself in the narrative world of a play that explores themes of revolution, the oppression of women, class politics and authoritarianism, by her own admission, therefore provides a key part of her ideological development. Significantly, in disrupting the division between herself and the character she plays, Vásquez Perdomo's account speaks directly to the narrative's transportational capacity to allow individuals to try out different identities and courses of action.
An understanding of transportation therefore gives theoretical weight to the notion that literary texts should be considered important by law enforcement practitioners when trying to understand the views and actions of terrorists. As demonstrated, narrative possesses a unique power and means to influence individuals' beliefs in a very different manner to rhetorical persuasion. Transportation is central in process by which terrorists weave an imaginary world around themselves and place the context of their actions within. In this sense, individuals not so much read literary texts but live them. Nonetheless, a series of questions arise in regard to many of those who engage in political violence: in particular, questions emerge as to whether the neat division between the 'real' and narrative worlds exist within these individuals' worldviews and what this means for attempting to understand their thinking.

Stories in context
Despite their importance, literary texts are only one means by which individuals come to encounter stories. From those historical and cultural narratives that underlie and form the basis of connection for entire peoples or nations to those told between friends or family members, stories are ubiquitous. Research that has focused on studying stories in the context of their telling has shown the value of this approach for explaining participation in social movements (Polletta, 2006;Polletta et al., 2011) as well as violence and crime (Fleetwood, 2015;Presser, 2009;Presser & Sandberg, 2015b). However, there are few studies of the flow of narratives in social spaces that explicitly link to participation in political violence. In this sense, it is crucial to understand how stories are composed, adopted, exaggerated, revised and/or rejected in various contexts and settings to better understand their influence. Stories, in particular, help bind collectives together including terrorist groups (Moore, 2010).
The study of autobiographical accounts then gives insights into where stories come from and what they do, as much as what they contain. Similarly, autobiographies themselves are necessarily performative (Porter Abbott, 2002, p. 135). These texts are often constructed by multiple authorsghost-writers, editors, publishersall of whom have the texts' eventual readers in mind. 6 Unlike terrorist communications, it cannot be argued that terrorist autobiographies, at some level, are not meant to entertain. However, as discussed, the idea that any narrative can be entirely one's own given the pervasive influence of culture is debatable. In this sense, even co-constructed autobiographies still offer significant insight into the worldviews of not only terrorists themselves but also the cultures and cultural settings within which they tell their stories. Nevertheless, analysts should always have the context of a story's telling in mind, whether this is the performative nature of autobiographical accounts themselves or the multitude of smaller stories contained within such texts.

Analytic framework
The second part of this paper seeks to provide a set of steps for bringing the theoretical insights discussed to bear on the analysis of terrorist authored accounts. Included is a framework for breaking down these texts into their constituent parts and then applying elements of narrative analysis directly onto them. This framework contains three levels for examining texts; paratexts; the story; and the narrative discourse. These elements can be further subdivided (see Figure 1.) Although in practice the elements discussed may overlap between these categories (most notably between story and narrative), the division is nevertheless used as an organising principle.
In order to demonstrate how this framework can be applied in practice, the accounts of two individuals who have engaged in political violence in different contexts will be explored. The first is the autobiography of Leila Khaled, a member of the Popular Front for the Liberation of Palestine (PFLP). Published in 1973, My People Shall Live, details Khaled's experience of being forcibly removed from her family's home in the city of Haifa to her role in the PFLP, including her involvement in the hijacking of two airliners. The second is the autobiography of Ommar Hammami (also known as Abu Mansuur al-Amriiki), an American who went to fight for al-Shabaab in Somalia. His account, The Story of an American Jihadi: Part One, was uploaded to the internet in 2012 and provides an in-depth look at his life growing up in the US and eventual journey overseas for jihad.

Paratexts
When approaching the analysis of a text, it is first beneficial to consider where exactly a narrative extends. Genette (1997) used the term 'paratexts' for this material, which lies somehow on the threshold of what would ordinarily be viewed as the narrative, or the presentation of events over time as contained within the text itself. Autobiography provides a good example of paratexts, given that the central textual narrative, or the selfaccount of the authors' life, is almost universally accompanied and contained within a body of paratexts. Commonly this includes titles, subtitles, artwork, photographs (and other less commonly included forms of images such as maps, drawings, etc.), prologues, forewords, and introductions. These texts form their own cultural worlds, in a sense. Even a cursory glance at the artwork that adorns the covers of such texts points towards something of a shared culturally-inflected notion of what a terrorist autobiography 'should' look like. For example, a number of autobiographies by individuals who participated in the conflict in Northern Ireland feature photographs or images of masked men armed with gunsoften with these weapons pointed directly at the reader (see for example -Black, 2008;Hamilton, 2008;Leslie, 2014). That such images dominate the covers of the accounts of individuals from both sides of the conflict may indicate a cultural understanding of participation in political violence, premising masculinity, that transcends the political divide.
Of particular interest, autobiographies often include prologues, which provide a space for authors to consider, discuss and reflect on the process of producing these self-narratives. As highlighted, Vásquez Perdomo, in the prologue of her autobiography, includes discussions about the process that speak closely to the conceptualisation of narrative and experience as inseparable (2005). In fact, relatively few terrorist autobiographies fail to contain some form of introspective assessment in their paratexts (former Ulster Volunteer Force operator David Hamilton's account (2008) being one notable exception).
Despite its apparently tangential status, all of this material can impact upon or even transform the light in which a narrative is read, without altering a single word (Porter Abbott, 2002, p. 26). In this sense, paratexts are therefore always part of the narrative, though they may however raise further questions about authorship and the co-construction of the autobiographical accounts. Nevertheless, even those accounts that are selfpublished and appear (in the sense of their construction and physical production at least) to be the work of a single author, usually follow a similar format, again highlighting common assumptions about the importance of paratexts in producing a proper account (see for example Eric Rudolph's memoir Between the Lines of Drift (2013)).

Content
Whilst this paper has been critical of the tendency of existing research to focus on the content of terrorists' accounts, at the expense of their form, this is not to dismiss the importance of this narrative element. Instead, literary theory provides a number of means to analyse the relative importance of events within a story that differentiate and demonstrate the benefits of narrative approaches from other forms of content analysis. Crucially, through narrative, individuals gather events into a meaningful story. Barthes and Chatman argue that a distinction can be made dividing these events into the constituent and supplementary (Barthes calling these 'nuclei' and 'catalyzers' (1982, pp. 295-296) whilst Chatman terms them 'kernels' and 'satellites' (1978)). Constituent events are those that cannot be removed without fundamentally changing the story; they are the key turning points and events driving the story forward (Barthes, 1982, p. 267). Supplementary events, by contrast, can be removed with the story remaining, on the face of it, intact. Whilst few studies draw explicitly on this analytic structure, much analysis of terrorist self-accounts implicit or explicitly presents hierarchical orderings of life events into similar categorisations of those deemed significant and those considered incidental (see for example Altier, Leonard Boyle, Shortland, & Horgan, 2017). However, it is clear that constituent events are only of greater importance than supplementary ones insofar as analysis is concerned with the sequence of events that comprise the story in itself. Yet, as this paper has sought to stress, there is more to a narrative than just story. The inclusion of these supposedly supplementary events raise questions just important as the inclusion of constituent ones; since they are not necessary to the story why did the author feel obliged to add them to the narrative? Asking this question is often advantageous in the interpretation of narrative, especially given that supplementary events often carry a significant burden of a narrative's meaning. The ability of narrative approaches to interrogate the minutiae of complex accounts is therefore useful in attending to how these smaller events give texture and establish meaning in a different, albeit sometimes complementary, manner to the essential elements of the story.
An equally profitable line of enquiry is to consider the events that were excluded from individuals' accounts. Whilst the approach advocated moves away from the necessity to compare self-accounts with events as objectively verifiable, there is nothing to say that narrative texts of terrorists be read in isolation from other accounts of these individuals' actions and lives. Here narrative approaches are distinct from other analyses of content. Absences in themselves are significant rather than simply a missing variable, and therefore pose questions as to why the author chose to exclude these elements from their accounts. An author may feel that it is unnecessary to include certain details from their accounts because they 'go without saying'or that a reader with knowledge of their narrative environment will already be aware of them. Sandberg argues that 'the most important stories in a society are often only hinted at, not fully told (2016, p. 166). Highlighting 'tropes', or common recurring literary and rhetorical devices that point towards or allude to familiar stories, in individuals' accounts is therefore a way to explore the dominant cultural influences within society (Sandberg, 2016 see also Cuddon & Preston, 1998, p. 948). In both of the senses explored above, narrative approaches allow the analyst to properly examine the importance of those stories that remain untold in terrorists' self-accounts.

Narrative discourse
Genre Whilst usually associated with works of fiction, self-accounts nevertheless consciously or unconsciously conform to and take on the shape and characteristics of genresor recurrent literary forms, such as epic, drama or tragedy. The genre to which a narrative text conforms is usually identifiable not only through its content but also via its tone. For example, Leila Khaled's autobiography, like dominant Palestinian narratives, is rooted in the genre of tragedy. Genres can also be further broken down into subgenres and narrative templates. For example, bildungsroman, or 'coming of age' stories focused on the psychological and moral growth of the protagonist from youth to adulthood, is one such template that terrorists commonly draw upon in the process of narrating their ideological and physical moves to extremist positions (Deary, 2010, p. 81). Similarly, structures that frame the author as 'the embattled, lone hero defying great odds, essentially moralself-righteously soin the face of a hostile society' are frequently employed (Aspeden & Haywoood, 2015, p. 253).
Relatedly, Joseph Campbell, in his classic work, The Hero with a Thousand Faces (1949), argues that all myths conform to the same underlying structure that he terms the 'Hero's Journey'. Omar Hammami's account (2012) of his journey from small town America to fight in Somalia with al-Shabaab, closely conforms to this narrative structure. Whilst Campbell's conceptualisation contains some 17 different stages, Hammami's description of outgrowing his hometown Salafi friends, who were not prepared to embrace and travel abroad for jihad, speaks directly to the idea of the 'call to adventure'. Here one's old ideals and settings are challenged as the protagonist becomes aware of a new and forbidden worldin Hammami's case the jihad being waged in Somalia. Similarly, Hammami reframes and shapes the way he surmounts the preventative efforts of his family and friends to stop him from travelling to Somalia as the first test on the 'road of trials' that he, as the hero, must overcome to achieve his goal. Reflecting back on his journey, Hammami himself specifically refers to the 'future trials' (2012, p. 100) his future self would have to face, whilst his declaration at one point that he 'walk from the south of Somalia to the north by foot if it would help matters' speaks directly to a narrative of trials that must be endured (2012, p. 42). Genres, sub-genres and other 'master-stories' should then be kept in mind when analysing these self-narratives. In particular, such considerations can help unearth how individuals both find meaning in their own lives as well as how they may wish the course of their life to be seen externally. This latter point is important given that autobiographies are necessarily performative (Porter Abbott, 2002, p. 135).

Time
Whilst autobiographies may appear at first glance to present a straightforward record of time,that is as an account of individuals' experiences as contained within the course of their livesthere are a number of temporal dimensions that the analyst should have in mind when approaching these texts. The first of these might be conceived of as the presentation of time within terrorists' accounts. Here, the structuralist distinction between story (the raw events of a story) and narrative discourse (these events as plotted and represented in a narrative) again proves informative, this time in demonstrating the temporal division at the heart of narrative. Story, with its progression of events through their chronological happenings in time is necessarily linear. Narrative discourse, on the other hand, is concerned with the form of time that reshapes that story in the telling. In other words, a textual account may jump between events in any order whilst still nevertheless maintaining coherence (Porter Abbott, 2002, p. 30).
In his classic work on narratology, Narrative Discourse, Genette (1980) presents three concepts for understanding the presentation of time: order, duration, and frequency. Here order is understood as the association between the order of events in the story and the alternative ordering of these same events as given in the narrative discourse. Duration refers to the amount of time dedicated in a narrative to narrate the events of the story as opposed to the time they 'actually' tookthe tempo or rhythm where the narrative speeds up and slows down key events in the story. For example, former al-Qaeda member Nasser al-Bahri's autobiography devotes just four pages to describing his life from birth to the age of 16by which time his commitment to jihad was apparently absolute (2013, pp. 12-15).
Here duration then helps craft a narrative that his participation was always beyond doubt. Finally, Genette uses frequency to describe the 'narrative temporality of repetition, which allows a narrative both to return multiple times to a single event and to condense multiple happenings of an event into a single instance of narration' (Martin, 2016, p. 3).
Genette's conception of these three elements forms a useful starting point for thinking about the temporal dimension contained within self-accounts. However, there also exists a more fundamental, if not complex and somewhat elusive, link between time and narrative. Ricoeur (1984, p. 52) has sought to reveal the existential or 'deeper experience of time' that narrative makes possible: Between the activity of narrating a story and the temporal character of human experience there exists a correlation that is not merely accidental but that presents a transcultural form of necessity. To put it another way, time becomes human to the extent that it is articulated through a narrative mode, and narrative attains its full meaning when it becomes a condition of temporal existence.
In other words, the true nature of timeas in terms more fundamental than of chorological units (such as days, minutes and hours)can only be rendered understandable when we narrativise it (Martin, 2016, p. 3). The accounts of terrorists therefore provide a window into their understandings of time. In doing so, they frequently reconfigure defining moments of history through the frame of their contemporary goals and concerns. Commonly, contemporary persecution and threats are framed as part of a wider historical subjugation, with significant incidents narrated to appear recent occurrences. In this sense, it is important to recognise that time is consistently shaped, stretched and compressed within terrorists' accounts, even if we initially think of these texts as straightforward representations of time. It is no coincidence that so many authors find themselves becoming involved in violence at 'breaking points' where historic persecution suddenly becomes unbearable. Framing their involvement in such a way lends authenticity, urgency and communal support for their actions. Khaled (1973, p. 24) provides a good example of this existential understanding of time in her autobiography. In describing her and her family's forced expulsion from their home, she states: I left Haifa four days after my fourth birthday, on April 13, 1948. My birthday was not celebrated because April 9 was a day of national mourning in Palestine. I am now twenty-nine years old and I have not celebrated a single birthday since, and will not do so until I return to Haifa. This understanding of time performs a number of functions for Khaled. First, this temporal, as well as physical, dislocation serves to fundamentally constitute what it is to be Palestinian. Second, this understanding of time provides a projective narrative, in the need to reclaim Haifa by whatever means, that 'imbues the time of individual lives with transcendent collective values' (Tölölyan, 1989, p. 101) and predicates the use of the use of violence in support of this cause. As such, to achieve the return to Haifa then is not only to repair the damage of the past and redeem the honour and dignity of her people, but also to 'restart time', and any prospect of a future, for all Palestinians.

Coherence
The search for a unified and coherent self-narrative remains the most common approach for examining life stories (Maruna, 2001, p. 7). However, the practice of mining narrative accounts for those events deemed most salient and significant, that has frequented the study of terrorists' accounts, presents the risk of smoothing out the author's experiences. It is crucial to recognise that a cohesive self-narrative may be the, sometimes necessary, product of the reductive analysis of the researcher. Furthermore, as existing research within narrative criminology has demonstrated, it may be more profitable for interpreting narratives to focus on ruptures or inconsistencies (Sandberg, 2013). In his analysis of Anders Breivik's manifesto, Sandberg (2013) demonstrates that the author 'utilizes several, sometimes competing, self-narratives or characters' (p. 75see also Bakhtin, 1981;Frank, 2012). These include the 'professional revolutionary'; 'the evangelist'; and 'the pragmatic conservative' but also the seemingly contradictory character of the 'social and likeable person', which Breivik develops through employing a relaxed tone, jokes and concerted efforts at humour (Sandberg, 2013, p. 78). Drawing on Jarvinen's argument that accounts are best understood if positioned against real or imagined accusations, 'Breivik's descriptions of himself as normal, likeable and social can be seen as a prepared response to accusations he knows he will meet, i.e. of being some kind of inhuman monster' (Sandberg, 2013, p. 79see also Jarvinen, 2003Jarvinen, , 2000. Omar Hammami is remarkable similar in deploying contradictory characters in his account. Whilst his personality as a light-hearted joker may be less affected than Breivik's efforts to portray himself as affable, Hammami is nevertheless keen to present himself as a 'social butterfly' (2012, p. 10). This character, nevertheless conflicts with an alternative character, of himself as hardened jihadi fighter (he reveals that he has authored a number of strategic documents for jihad under a pseudonym within the opening pages), which he deploys at other points in his account. Again, in a similar vein to Breivik, Hammami also appears to be using this lighter character to defend himself from accusations, namely that his decision to travel to Somalia was the result of him being something of a loner or someone marginalised in America. As such, he states 'I left America while the American dream was firmly tucked under my pillow ' (2012, p. 113).

Conclusion
This paper has sought to demonstrate that a better appreciation of narrative is critical for understanding terrorism, in particular the experiences of those who have engaged in political violence. As such, a theoretical basis and methodological framework for studying terrorists' self-accounts as narrative texts in their own right has been presented. In doing so, this paper has touched upon, and at points attempted to address, a number of existing problems regarding the use of narrative within the broad field of terrorism studies.
The term narrative has been used uncritically and without care in various aspects of research on terrorism. Similarly, whilst there is growing recognition of the utility of studying terrorist-authored accounts as narrative texts, there appears to be some hesitancy on behalf of analysts in the field of terrorism studies to enter into proper discussions of the nitty-gritty of how to best approach and interrogate them. The idea that these accounts 'speak for themselves' and the dominant position of psychological approaches, that are premised on the idea that individuals ultimately seek coherence in giving accounts of their lives, may both have had an impact here. However, within literary studies and narrative criminology there are innovative methodological approaches, some of which directly challenge the smoothing out of individuals' life stories, that can be brought to bear and used in conjunction with psychological approaches. A greater appreciation of the means and methods of narrative analysis and narrative theory are necessary to properly utilise the distinct features of this approach. It may appear at first somewhat counterintuitive to focus on what is excluded rather than included from an account or examine the presentation of time in these texts cut free from their chronological basis but it is precisely here that narrative analysis offers unique insight. As such, whilst these elements are often difficult to examine methodically, they are no less than crucial for better comprehending how individuals involved in militancy understand the world, draw upon existing narrative resources and give meaning to their actions.
Returning to these methodological concerns, it is hoped that the framework presented in this paper goes some way in forming a basis for the systematic application of various elements of narrative theory. As alluded to, further work is needed to defend narrative approaches of the study of terrorism from criticisms of their vagueness, superficiality and subjectivity. Demonstrating a clear, consistent and robust means for utilising these tools is therefore a critical and ongoing task. Given that this paper advocates a holistic approach to the place of narrative in understanding involvement in terrorism, it is hoped that the utility of this framework for examining stories in context is clear. As emphasised, it is crucial to examine when, why and how stories come to be told, heard and the resources that narrators draw upon in telling them. In this sense, autobiography gives a unique window into the multitude of smaller stories that individuals encounter over the course of their lives, as well as the larger accounts they aim to give of their own experiences. Analysing these smaller stories may then be extremely profitable. Here the author may cross between and be both storyteller and recipient of the same or different stories. As highlighted, terrorists are both interested and influenced by stories, and are storytellers themselves. Here it is crucial to have in mind the question: what purposes do stories, and the act of storytelling do? Whilst there is recognition that storytelling may serve an important function as a cultural practice in binding members of terrorist groups together (see, for example, Hegghammer, 2017, pp. 191-192), there has been little systematic study of the sorts of stories that are told or where and when these tellings take place. Similarly, examining how certain stories are adopted, revised, exaggerated or discarded within the milieus that surround terrorist groups, and from which they draw, may help analysts understand how participation in political violence comes to be viewed as attractive or necessary. In particular, this may include stories that are not necessarily related to a group's ideological beliefs but rather those that provide other reasons for involvement, for example standing with one's peers or a sense of duty. In this sense, it is hoped that the flexibility of this methodological framework may prove useful for research that focuses on what do stories do as much as what they reveal.
Finally, it is hoped that this paper will provoke further discussion regarding the worlds that terrorists create, weave around themselves and place their actions within. Here it may be somewhat misleading to use the term 'imaginary worlds', given how theoretical insights from literary studies openly demonstrate the real-world impact these processes can have. Such research holds potential benefits for those law enforcement practitioners tasked with understanding why individuals have engaged in violence. Crucially, it is clear that literary texts should be afforded significant weight in the course of such investigations given their capacity to influence individuals' beliefs in a uniquely emotive manner. Notes 1. For some of these academic works within terrorism studies see Tölölyan (1989); Halverson, Goodall, and Corman (2011); Ramsay and Marsden (2013); Braddock and Horgan (2016). Examples of those works outside of the field of terrorism studies, but of direct relevance include Bearman and Stovel (2000), Blee (2002Blee ( , 2005 and Polletta (2006) in Sociology; Moore (2010) in Political Science; Ringmar (1996) in International Relations; Colvin (2009) in German Studies;andSandberg (2013, 2015) in Criminology. 2. Ramsay and Marsden (2013) and Braddock and Horgan (2016) both draw upon relevant theoretical literature on narrative within literary studies. However, these scholars remain heavily influenced by psychological accounts of narrative, in particular, the work of McAdams (1988McAdams ( , 1993. Unlike other psychologists who work on narrative, such as Bruner (1990) or Polkinghorne (1988), McAdams makes little reference to any literary theory of narrative, instead drawing on Erik H. Erikson's work on identity (1968,1975). 3. For example, the story/narrative discourse distinction is useful for beginning to think about how time is presented within terrorist accounts. Nevertheless, it is useful to demonstrate how in this understanding, 'story' and 'narrative' are inherently bounded by different constraints. A story is necessarily linear and can only present events in their chronological order in the direction time moves. By contrast, narrative discourse can jump between these events in any order and still maintain coherence (Porter Abbott, 2002, p. 30). 4. John Morrison has floated the idea of studying terrorists' mental attack preparations through the lens of 'positive visualisation' as widely used in sports science (2017). Similarly, Anders Breivik dedicates time within his manifesto to imagining the execution his attack, complete with musical accompaniment; I simulate/meditate while I go for a walk, playing my Ipod in my neighbourhood. This consists of a daily 40 minute walk while at the same time philosophising ideologically/performing self indoctrination and the mental simulation of the operation while listening to motivational and inspiring music (2013, p. 854).
5. The perception of narratives' power to resonate emotionally can be further demonstrated by the actions of totalitarian governments who have often sought to suppress even fictional narratives given fears of their potential role in helping enact social change (Green & Brock, 2002, p. 315). 6. Defining the parameters of autobiography is a difficult undertaking. As Lynch argues, 'autobiography is arguably the most underhand of all literary genres, consistently avoiding the definitions fashioned for it and eluding the genre boundaries expected of it' (Lynch, 2009, p. 1). A broad but workable definition of such texts is commonly taken to be 'a book understood by its author, its publisher, and its readers to be a factual account of an author's life' (Yagoda, 2009, p. 1).